Deep Q-Learning for Directed Acyclic Graph Generation

TL;DR

This paper introduces a deep reinforcement learning approach using deep Q-learning to generate directed acyclic graphs with specific structures and node types, addressing a challenging graph generation problem.

Contribution

The paper presents a novel deep Q-learning method tailored for generating directed acyclic graphs with desired topologies and node types, a task not well addressed by existing methods.

Findings

Successfully generates DAGs with specified structures

Operates effectively in sparse reward environments

Demonstrates applicability across various graph generation scenarios

Abstract

We present a method to generate directed acyclic graphs (DAGs) using deep reinforcement learning, specifically deep Q-learning. Generating graphs with specified structures is an important and challenging task in various application fields, however most current graph generation methods produce graphs with undirected edges. We demonstrate that this method is capable of generating DAGs with topology and node types satisfying specified criteria in highly sparse reward environments.